[mythtv] OSX backend obstacles?

David Abrahams dave at boost-consulting.com
Sun Dec 18 14:06:38 EST 2005

Daniel Kristjansson <danielk at cuymedia.net> writes:

> On Sun, 2005-12-11 at 13:54 -0600, Robert Johnston wrote:
>> Would it be possible to "Modularize" the Myth calls for recorders, so
>> you have an IVTV Module, a Firewire Module, a DVB Module and so on?
>> This would make it pretty easy to make a "Linux Firewire Module" and
>> an "OSX Firewire Module" (And, say, a "Solaris Firewire Module" or
>> whatever else people want to do) and to dynamically link them in at
>> runtime (So if you get a Firewire HD box, you don't need to completely
>> recompile Myth to use it, just compile the Required Firewire module,
>> drop it in, and restart MythBackend).
> Most of the recorders are different enough that they need special
> cases in tv_rec. But there are plenty of modularized portions, i.e.
> the RecorderBase and DTVRecorder classes. For firewire, you wouldn't
> want to create a recorder, it's just a half dozen calls that need
> to be encapsulated, an OSX programmer could probably knock the port
> off in an afternoon.

Easier said than done, it turns out.  Well, I'm not an OSX programmer
really, but the ways to access firewire video on OSX are myriad, and
not well documented.  Even an experienced OSX programmer wouldn't have
an easy time of it unless he were intimately familiar with this part
of the API.  I can capture it to a QuickTime file easily enough, but
the innards of MythTv seem to want mpeg-2 frames (?) and it's not
obvious how to deliver those using MacOS.  It's not even easy to
understand what the mythtv code is doing, since libiec61883 isn't
documented in any place that I could find.

Dave Abrahams
Boost Consulting

More information about the mythtv-dev mailing list